Posts Tagged policy

How I Missed The 6 Hour Pneumonia Antibiotics Window

When you come in morbidly obese, in this corner, weighing in at 350+ pounds and your physician is having to confirm that you will not break the CT scanner, let’s all just admit that everything gets a lot harder: making diagnoses, finding veins, dosing medications.

When you’re a nonsmoker, nonasthmatic morbidly obese person who comes in short of breath with leg swelling for the past several days, you’re on Lasix at home, and you’ve got some wheezing in your gigantic lung fields, and your chest x-ray looks like pulmonary edema, your doctor thinks he’s made the diagnosis.

But you’re very hypoxic — O2 sat in the 70s on room air — even after 6 hours in the ED, after lasix and some nitroglycerin. So we scan you for pulmonary embolus, worried about a PE. You have no PE, not really much pulmonary edema, but you have evidence of pulmonary hypertension (hi, obstructive sleep apnea) and a small consolidation, even without cough, or fever rectally, we hang your classic ceftriaxone/azithromycin. Missed that all important “6 hour window.”

And thus, our rant begineth.

Similar to my medical errors rant, I think a lot of emergency physicians have problems with these guidelines, which are described as quality indicators and let the public evaluate a hospital based on these guidelines. This assumes that a score of “100%” is the absolute best score for a hospital. So, batter up:

  1. Hey! Medicare! These are guidelines. Not rules. These are to help us physicians guide our therapy, not to control it for us. Individual patients come with individual problems that cannot always be boiled down in a document.
  2. Give humans (in this case, physicians) a perverse incentive, and we’ll start acting perverse. This can go in two ways.
    • Don’t want to be dinged for not giving antibiotics on time? Admit the patient with a diagnosis of “shortness of breath” instead of “pneumonia.” If you’re not in the inclusion criteria, you’ll sneak right by. (I’m not suggesting that physicians actually do this in practice, just giving an example.)
    • Want to make sure you meet those all-important guidelines? Maybe there’s an “early pneumonia,” or the diaphragm’s a little hazy on a portable film? Just give antibiotics to cover your ass (and your hospital’s), even though it might not be what’s best for the patient (example: the patient with hyponatremia who just last week finished a 2-month long battle with C. diff.)
  3. Scientific evidence indicates that the following process of care measures represent the best practices for the treatment of community-acquired pneumonia. Higher scores are better. Okay, so, what’s the data say?
    • One of the leading advocates of this is Dr. Peter Houck, who’s done a bunch of research showing better outcomes with early antibiotic administration. The problem? It’s all retrospective, data-mining from large data sets. (A huge slide deck from Dr. Houck from 2006 provides some rebuttals to this argument.) The data also shows that there’s a difference between antibiotics at before versus after 8 hours; perhaps the magic 6 hour window is a compromise?
    • Another “quality measure” is blood cultures before antibiotics given. And this one is simply just foolish. Antibiotics for pneumonia rarely if ever change clinical practice. This has been shown in multiple studies, from the pulmonology literature to the British Emergency Medicine literature (“30 (1.4% of all cultures) were “true positives” and 4 (0.18%) influenced subsequent patient management.”). Also multiple studies in our own Emergency Medicine journals refuse the need for cultures, too. They rarely, if ever change clinical management, yet they’re “quality indicators.” We might as well have a guideline to order ESR/CRP on patients with suspected pneumonia, too. Give me a break.
  4. And finally, the concept of an acceptable miss rate is — unacceptably — missing from the discussion. Like the general surgeon who misses a few appies or who removes a few normal ones, we should be wary of anyone that reports or preaches 100% compliance to some of these guidelines: are these physicians thinking about risks and benefits and weighing options, or just blindly following? Should we aim toward always getting things right 100% of the time? Absolutely. But the real and theoretical worlds collide. There probably is some small benefit for early antibiotics, and most people most of the time should get them earlier rather than later. (We already have incentives to do this: they can leave the ED faster and move to the floor!) But there should always be a small percentage of cases that don’t fit inside the 95% confidence interval (usually around 5% of them): a group of people who live outside the standard deviations. Aspirin for an MI? Almost always. But how about the MI with the GI bleed with the hematocrit of 10? Or the patient with the anaphylactic aspirin allergy? Risk, benefit. No right answer.

, , , ,

2 Comments

Fixing Medical Err-ERs

The NYT has an Op-Ed by a former head of the National Transportation Safety Board talking about reforming the health care system and reducing medical errors. He cites the To Err Is Human Institute of Medicine report suggesting 98,000 annual deaths and billions of dollars due to medical errors, and notes:

What makes the problem all the more frustrating is that we could address it with little cost to the American taxpayer. Because American medicine accepts error as an inevitable consequence of treatment, our hospitals, insurers and government do little to respond to unnecessary deaths. If we are to address the problem in a serious manner, we must first change this culture.

He then goes on to start talking about “Hospital-acquired infections, for instance, affect millions of Americans each year,” and loses me, after comparing “our roads, rails and skies” to our medical system.

Don’t get me wrong. I’m the first to admit that the medical system makes errors. “To err is human,” and I can’t help but agree. As physicians we’re doing our best to simplify incredibly complex biological systems down to a level that our brains can comprehend and try to evaluate and treat. We also have good days and bad ones, and on occasion even let our emotions affect our logical, rational, objective decision making. Damn right I accept errors as a consequence of treatment. I don’t like them; I would like to minimize them, but I accept them. Everything in life is risk-benefit.

So there’s point number one: we’re human, and by definition, flawed. Doctors are not gods, do not know everything, do not have all the answers. And we will, sadly, make medical errors.

Number two: what’s a medical error? The report this whole thing is based on says there’s a bunch of types, but how many are really fixable?

  • Diagnostic Error. ex: Error or delay in diagnosis
  • Treatment Error. ex: Error in the dose or method of using a drug
  • Preventive Error. ex: Failure to provide prophylactic treatment
  • Other Errors. ex: Failure of communication, Equipment failure

Medication errors are often the big, obvious, low-hanging fruit. It’s why we can’t write “MgSO4” for magnesium, because our crappy handwriting sometimes was interpreted as “morphine sulfate.” It’s things like the Dennis Quaid heparin saga when his neonate twins were given adult doses of heparin (adult dose heparin should probably be a different size or color and should probably never be available on a pediatric floor, duh). But other errors are much harder to fix, and probably come with even more unintended consequences (see #4).

Number three: can we really simplify the medical system down to levels that they can in the transportation industry? Sure, there are plenty of things that can go wrong with a plane, internal and external factors like engine malfunctions and torrential rains, but I feel like there’s many more that can go wrong with humans–not just the patient, but also the multiple humans taking care of the patient. Yes, there’s definitely obvious areas like central line placement checklists and the anesthesia world learning from aviation standards and checklists, but I can’t imagine any other place being more challenging than the barely-controlled chaos of a busy emergency department.

And finally for the ultimate problem in de-bugging medical errors: unintended consequences. This is the art of public policy where things often go awry. Take, for example, “medical errors” of Medicare’s “never events”:

  • No Serious Disability from Falls. Lofty goal, sounds great, right? But what do we do with the elderly demented patient who continually gets out of bed? Restrain them, sedate them, or have someone watch them 24/7? Falls are an awful thing, and I’ve seen first hand the consequences, but there are plenty of consequences to preventing every single one of them in every single patient, too. What about the young patient who is told not to walk on their broken leg, but still believes they’re strong enough? Do we sedate or restrain everyone who has the potential to fall?
  • No Catheter-Associated Urinary Tract Infections. Again, sounds good, but can we really get this down to zero in the intubated, diabetic, HIV, cancer patient in CHF who needs monitoring of their I’s and O’s? What’s the solution to this “never event,” cipro-coated catheters? (And then for the now-cipro-resistant bugs we’ve just created?)
  • No DVT/PE after Orthopedic Procedures. So the hypercoaguable patients just won’t get orthopedic procedures, because orthopods just won’t touch them anymore, for risk of triggering a “never event,” which the hospital won’t pay for.

I could go on, but you get the point. These events should be “almost-never events.” Medicare has plenty of data for all this stuff; they should come up with a reasonable standard that every hospital should have to meet. “Average percentage of UTIs from foleys today in a 500 bed hospital: 3%. All similar hospitals must all now be under 3%.”

In nowhere are these tradeoffs more apparent than in the Emergency Department, where everything is probably just a liiiittle more error prone, with multiple sick patients, little to no medical history on patients, and a limited time frame in which to act: diagnose, treat, and dispo patients. Antibiotics for pneumonia in 6 hours? Great for the obvious febrile coughing patient with infiltrate. But add anything else in there, and the story gets murky, while we still get judged on our “quality.” To give or not to give, which is the medical error?

I’m all for Jim Hall coming to visit my Emergency Department to figure out how I can do better for my patients, as long as he realizes that there’s no free lunch, everything has a risk and a benefit, and that at some point, in medicine minimizing one error just maximizes another.

, , ,

1 Comment